Model comparison
gemma-3-4b-it-roleplay-tuned-v2 vs Qwen3-VL-8B-Instruct-abliterated-v1
At Q4_K_M, gemma-3-4b-it-roleplay-tuned-v2 is the smaller download — 2,489,894,848 bytes against 5,027,785,792.
From the file· summed bytes, KV per layer
Side by side
| gemma-3-4b-it-roleplay-tuned-v2 | Qwen3-VL-8B-Instruct-abliterated-v1 | |
|---|---|---|
| Parameters | 4.3B | 8.8B |
| Architecture | gemma3 | qwen3vl |
| Layers | 34 | — |
| Native context | 131,072 | — |
| Mixture of experts | no | no |
| Quantizations published | 36 | 48 |
| Smallest quantization | 1.06 GiB | 1.97 GiB |
| Q4_K_M | 2.32 GiB | 4.68 GiB |
| Licence | gemma | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | gemma-3-4b-it-roleplay-tuned-v2 | Qwen3-VL-8B-Instruct-abliterated-v1 | Ratio |
|---|---|---|---|
| 4,096 | 0.25 GiB | — | — |
| 8,192 | 0.33 GiB | — | — |
| 16,384 | 0.48 GiB | — | — |
| 32,768 | 0.79 GiB | — | — |
| 65,536 | 1.42 GiB | — | — |
| 131,072 | 2.67 GiB | — | — |